Predicting Upcoming Board Games

Predictive Models for BoardGameGeek Ratings

Pipeline

targets::tar_visnetwork(targets_only =T)

Models

model_board = pins::board_folder("models",
                                 versioned = T)

averageweight_fit = 
    vetiver_pin_read(
        model_board,
        "bgg_averageweight"
    )

average_fit = 
    vetiver_pin_read(
        model_board,
        "bgg_average"
    )

usersrated_fit = 
    vetiver_pin_read(
        model_board,
        "bgg_usersrated"
    )

Assessment

valid_predictions |>
    pivot_outcomes() |>
    left_join(
        games |>
            bggUtils:::unnest_outcomes() |>
            select(game_id, usersrated),
        by = join_by(game_id)
    ) |>
    plot_predictions(alpha = usersrated)+
    theme(legend.title = element_text())

targets_tracking_details(metrics = valid_metrics,
                         details = details) |>
    select(model, minratings, outcome, any_of(c("rmse", "mae", "mape", "rsq", "ccc"))) |>
    filter(minratings == 25) |>
    select(minratings, everything()) |>
    gt::gt() |>
    gt::tab_options(quarto.disable_processing = T) |>
    gtExtras::gt_theme_espn()

minratings	model	outcome	rmse	mae	mape	rsq	ccc
25	lightgbm	average	0.692	0.509	7.582	0.282	0.463
25	lightgbm	averageweight	0.457	0.347	19.210	0.665	0.804
25	lightgbm+lightgbm	bayesaverage	0.293	0.170	2.823	0.432	0.646
25	lightgbm	usersrated	1565.092	461.642	184.381	0.229	0.476

Predictions

predictions = 
    upcoming_games |>
    impute_averageweight(
        model = averageweight_fit
    ) |>
    predict_bayesaverage(
        average_model = average_fit,
        usersrated_model = usersrated_fit
    )

predictions |>
    filter(yearpublished >= 2024) |>
    # this goddamn bah humbug game
    filter(game_id != 388225) |>
    predictions_dt(games = games) |>
    add_colors()